Interstage trafo in vari-mu circuits

While I understand why its there, like in a UA175, I have no idea about the common winding ratios used in these design.

I think I´ve read somewhere thats its usually a step-down trafo from VCA plates to output amp grids. Say maybe 2:1 or 4:1.

But I cant remember where I sourced that information and Im not sure why one couldnt use a 1:1 ratio. Like a common 10K:10K.

Anybody has any thoughts on this?

Thanks.

There's really no "rule." You can use a step-down, step-up or 1:1 depending on the requirements of the circuit.

There's a few issues involved with selecting/using an interstage transformer in a vari-mu compressor, but these are outside the scope of your question.

Thanks.

There's a few issues involved with selecting/using an interstage transformer in a vari-mu compressor, but these are outside the scope of your question.

Click to expand...

No please. Do go on.

The only issue I can think of would be that of DC unbalance between the plates during GR and that might call for a airgapped core.

I dont know how much unbalance is to be expected, but if it is not to much Im thinking a cheap ungapped Edcor could do the job.

search all the various threads; there are substantial benefits gained from use of interstage transformer after variable mu stage. You can do it without, but not as well.

I have used a general purpose 1:1 matching transformer as a vari-mu interstage. The trick is to manage the source and load impedances carefully.

By way of illustration, here's a set of transfer curves for the venerable 6386:



Suppose you're using a 10K:10K interstage with regulated B+ fed to the plates via the primary centertap. The plate supply voltage will be essentially constant, but the mu, Gm and Rp vary substantially as the plate current is varied.

Look at what happens to Rp as the plate current is reduced! Now, imagine a 10K xfmr primary fed from a source impedance that can vary in an instant from 4K up to 20K or more. What do you suppose happens then?

To deal with this, you can use a special transformer with a massive primary inductance, or provide a means to fix the source impedance to a value low enough to swamp out variations in Rp.

You can figure out the specifics if you think about it long enough. Or just search around some more.

You want a compressor to be fast, what happens when you put iron and copper in the way?

check the buffer on the 670 control input.

check out the Langevin Level Line

check reaponse time of 175 vs leveline

.

[quote author="CJ"]You want a compressor to be fast, what happens when you put iron and copper in the way?[/quote]

What happens when you charge/discharge capacitors through large resistances?

My compressor uses an interstage and has an attack time of 500uS.

I would bet you were using a 1:1
Yes on the time constants.
There is a whole AES on that subject alone.
An RC for bass, an RC for cymbels, and an RC for all points in between.

The biggest issue with the interstage in a mu-compressor is simply that it is a hi-z to hi-z device which encounters the worst effects of stray capacitance and inductance. So anything less that a reaaly really good transformer will not be flat enough.

Dave I like your new avatar.




So much I, had it tattooed onto my arm.




Actually, I did that about 5 years ago, to celebrate my gearhead, Detroit origins, lol.
But it's cool that you made it your avatar!

You guys ARE SOOOO BUSTED!

:grin: :razz: :twisted: :grin: :green: :twisted: :razz: :grin:

sAY IT AIN'T SO!

aND THE TWOEOPLE WHO HATE ME THE MOST.
iT ALL MAKES SENSE NOW.
:razz:

wHO ELSE IS WEARING THAT SECRET ARM BAND?
jENNS AND kUBI i BET.
:shock:

[quote author="Kit"]I dont know how much unbalance is to be expected, but if it is not to much Im thinking a cheap ungapped Edcor could do the job.[/quote]

Beat you to that idea years ago. It's faster than the garden variety mu compressors too. This is the controversial Poor Man's 660. It crushes, man.

analag

*Hand to side of mouth* When we say "mu", we really mean transconductance. But it has 16 letters.

More like transcontenental, or transvestite ...
:razz:
I thought it wasa transmuconductance...

but this aint jepeordy, and I sure the hell aint alex trebec.

boy was i wrong.

i enter transmuconductance into the goolge engine without quotes, and it turns up absolutely nothing.
probably the first time for that.

try it yourself if you no believe.

We aren't changing the mu with the control grid voltage, we are changing the transconductance.

And copper and iron don't slow you down, high-impedance time constants that don't snap to attention like a fuel dragster do. Cause they aren't high impedance until the charge cycle. Til then, you better have some juice to get off the line.

Torque.

Yeah, low dcr,
hey speaking of xfmr;s,
hows this sound for the WE pre?
see if you can decipher the idea from the pic for five bonus points:

I see the series-resonant circuits in there.
But how is feedback coupled from that node?
That's not clear in the drawing.

The idea is to break down a 9:1 xfmr into a couple of 3:1 xfmrs, as to avoid all the angle from the leakage.
Thus, avoiding instability in the amplifier due to feedback being out of sync with input.

If there was any residual angle left after the scheme above, you could fine tune it, kind of like a hammie build out of an IF stage, if you were ever into stuff above 20 kc and I know you were.

Once you tap into a coil, you are kind of stuck with that ratio and that angle.
This puts control of the feedback into an easy to get to area for tweaking. Rebuilding a winding every hour to adjust the taps could be a real bummer.

This opens up a rotary switch for variable feedback amplitude and angle adjust as a possible front panel control.

You might not need the build out filters, and I do not even know what kind of filter would work best, it could be fine tuned on the outside, then stuff the whole enchilad into a mu can and seal it up so nobody knows how thw F you got a 9:1 output to run so flat, get me?

Hey CJ, send some of what you're smokin! I need some here! That's a wack idea. I'm not smart enough ATM to figure out if you're on something (well, I know you're on something) or not, but I find it intriguing. Keep it up!

Reminds me of the webs made by the spiders they fed LSD to, CJ, but I get it.

The mid-access point is good for tweaking, but it might make the feedback pick-off load-sensitive to the whole thing.
More so than a tickler-coil on the side would be.
I would have to terminate it at the cathode end correctly.
I probably need low-z, too. A few k at least.